I'm tring to prove that the unit circle $$S^1=\{(x_1,x_2)\in\mathbb{R}^2\text{ such that }x_1^2+x_2^2=1\}$$ is an embedded submanifold of $\mathbb{R}^2$ using the following Characterization:

A nonempty subset $M \subset\mathbb{R}^n$ is an m-manifold iff:

For every $p\in M$, there are two open sets $O,W\subset\mathbb{R}^n$ with $0_n\in O$ and $p ∈ M ∩ W$, and a smooth diffeomorphism $ϕ: O → W$, such that $ϕ(0_n) = p$ and $$ϕ(O ∩ (\mathbb{R}^m × {0_{n−m}})) = M ∩ W$$.

My attempt:

lets fix $a\in S^1$, there exists a unique $\theta_0\in[0,2\pi)$ such that $a=(cos(\theta_0),sin(\theta_0))$

and let $\phi$ be the diffeomorphism \begin{array}{cccc} \phi : & (0,\infty)\times(\theta_0-\pi,\theta_0+\pi) & \longrightarrow & \mathbb{R}^2\backslash D_{\theta_0+\pi}\\ ~~ & (r,\theta) & \mapsto &(rcos(\theta),rsin(\theta)) \end{array}

where $D_{\theta_0+\pi}$ is the half line at the origin with polar angle $\theta_0$

From here i dont know how to proceed. Should i modify this map or use as it is? and who are the two open sets of the charactrization?


The way you constructed your $\phi$ is almost good. Intuitively, the characterization you presented requires $\phi$ to be a diffeomorphism between 2 open sets $W$ - the ambient space around $\mathbb S^1$, and $O$ - a deformed version of $W$ such that $\mathbb S^1 \cap W$ is "squished and flattened" into $\mathbb R \times \left \{ 0 \right \}$ (by $\phi^{-1}$). You could think of $\phi|_{O\cap (\mathbb R \times \left \{ 0 \right \})}$ as a parameterization of the circle using $\theta$.

Your $\phi$ is a good candidate besides the fact there needs to be some alignment so that $0$ is mapped to $a$: substituting $\theta+\theta_0$ into the trig functions instead of $\theta$, and using a radius $1+r$ instead of $r$. This way we get $W=\mathbb R^2 \setminus D_{\theta_0+\pi}$ as you said but $$O=(-1,\infty)\times(-\pi,\pi)$$ From the naturality of polar coordinates it's pretty easy to convince $\phi$ is bijective, two points can't have the same radius and angle, and every point has a radius and an angle. $\phi$ is also smooth due to its components being elementary functions. Hence, it is a diffeomorphism.

$\phi(O\cap (\mathbb R \times \left \{ 0 \right \}))\subseteq \mathbb S^1 \cap W$ since any point of the form $(0,\theta)$ is mapped to the unit circle at an angle which isn't $\theta_0+\pi$, and $\mathbb S^1 \cap W \subseteq\phi(O\cap (\mathbb R \times \left \{ 0 \right \})) $ using the reverse argument.

  • $\begingroup$ thank you sir you're a genius So we chose our diffeomorphism to be \begin{array}{cccc} \phi : & (-1,\infty)\times(-\pi,+\pi) & \longrightarrow & \mathbb{R}^2\backslash D_{\theta_0+\pi}\\ ~~ & (r,\theta) & \mapsto &((1+r)cos(\theta+\theta_0),(1+r)sin(\theta+\theta_0)) \end{array} $O=(-1,\infty)\times(-\pi,+\pi)$ and $w=\mathbb{R}^2\backslash D_{\theta_0+\pi}$ $\endgroup$ – Donnie Darko Oct 23 '20 at 18:28
  • 1
    $\begingroup$ Yup, I think that should work @DonnieDarko $\endgroup$ – Theorem Oct 23 '20 at 19:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.